Department of Medical Oncology, 2 Hospital of Zhejiang University College of Medicine, Hangzhou 310009, Zhejiang Province, China.
World J Gastroenterol. 2013 May 7;19(17):2650-9. doi: 10.3748/wjg.v19.i17.2650.
To develop a prognostic model to predict survival of patients with colorectal cancer (CRC).
Survival data of 837 CRC patients undergoing surgery between 1996 and 2006 were collected and analyzed by univariate analysis and Cox proportional hazard regression model to reveal the prognostic factors for CRC. All data were recorded using a standard data form and analyzed using SPSS version 18.0 (SPSS, Chicago, IL, United States). Survival curves were calculated by the Kaplan-Meier method. The log rank test was used to assess differences in survival. Univariate hazard ratios and significant and independent predictors of disease-specific survival and were identified by Cox proportional hazard analysis. The stepwise procedure was set to a threshold of 0.05. Statistical significance was defined as P < 0.05.
The survival rate was 74% at 3 years and 68% at 5 years. The results of univariate analysis suggested age, preoperative obstruction, serum carcinoembryonic antigen level at diagnosis, status of resection, tumor size, histological grade, pathological type, lymphovascular invasion, invasion of adjacent organs, and tumor node metastasis (TNM) staging were positive prognostic factors (P < 0.05). Lymph node ratio (LNR) was also a strong prognostic factor in stage III CRC (P < 0.0001). We divided 341 stage III patients into three groups according to LNR values (LNR1, LNR ≤ 0.33, n = 211; LNR2, LNR 0.34-0.66, n = 76; and LNR3, LNR ≥ 0.67, n = 54). Univariate analysis showed a significant statistical difference in 3-year survival among these groups: LNR1, 73%; LNR2, 55%; and LNR3, 42% (P < 0.0001). The multivariate analysis results showed that histological grade, depth of bowel wall invasion, and number of metastatic lymph nodes were the most important prognostic factors for CRC if we did not consider the interaction of the TNM staging system (P < 0.05). When the TNM staging was taken into account, histological grade lost its statistical significance, while the specific TNM staging system showed a statistically significant difference (P < 0.0001).
The overall survival of CRC patients has improved between 1996 and 2006. LNR is a powerful factor for estimating the survival of stage III CRC patients.
建立预测结直肠癌（CRC）患者生存的预后模型。
收集 1996 年至 2006 年间接受手术的 837 例 CRC 患者的生存数据，通过单因素分析和 Cox 比例风险回归模型进行分析，以揭示 CRC 的预后因素。所有数据均使用标准数据表记录，并使用 SPSS 版本 18.0（SPSS，美国伊利诺伊州芝加哥）进行分析。通过 Kaplan-Meier 方法计算生存曲线。采用对数秩检验评估生存差异。通过 Cox 比例风险分析确定疾病特异性生存的单因素风险比和显著且独立的预测因素。逐步程序的阈值设定为 0.05。定义 P < 0.05 为统计学意义。
3 年生存率为 74%，5 年生存率为 68%。单因素分析结果表明，年龄、术前梗阻、诊断时血清癌胚抗原水平、切除状态、肿瘤大小、组织学分级、病理类型、脉管侵犯、邻近器官侵犯和肿瘤淋巴结转移（TNM）分期是阳性预后因素（P < 0.05）。淋巴结比值（LNR）也是 III 期 CRC 的强预后因素（P < 0.0001）。我们根据 LNR 值将 341 例 III 期患者分为三组：LNR1，LNR ≤ 0.33，n = 211；LNR2，LNR 0.34-0.66，n = 76；LNR3，LNR ≥ 0.67，n = 54。单因素分析显示，这三组患者 3 年生存率有显著统计学差异：LNR1，73%；LNR2，55%；LNR3，42%（P < 0.0001）。多因素分析结果显示，如果不考虑 TNM 分期系统的相互作用，组织学分级、肠壁浸润深度和转移淋巴结数量是 CRC 最重要的预后因素（P < 0.05）。当考虑 TNM 分期时，组织学分级失去统计学意义，而特定的 TNM 分期系统显示出统计学差异（P < 0.0001）。
1996 年至 2006 年间，CRC 患者的总体生存率有所提高。LNR 是评估 III 期 CRC 患者生存的有力因素。
